55 comments

[ 3.3 ms ] story [ 80.2 ms ] thread
Hm, isn't it a risk to pay too much?

If you've got enough money that you can't spend it in a lifetime, why continue to do extremely difficult work?

800k will be around 400k post tax. Nowhere near a life changing sum in bay area.
Well, first off is not even hard to spend 800k

Second, many smart people like to work.

> Well, first off is not even hard to spend 800k

If you're buying a house, sure, but that's just the 1st year.

> ..many smart people like to work.

I do too, but life outside of work, being financially independent - that's tempting.

You're not going to net anywhere close to 800K. As the article notes, they're talking 300K base + 500K equity (which is illiquid) at the 'highest end.' (In reality the highest end is likely quite a bit higher, but with a wide and thinly distributed band).

- ... the 300K base is 177K after tax.

- ... the 500K equity is ~250K after tax.

Financial independence assuming the stock doesn't massively appreciate by the time you see liquidity is 5-10 years away. I mean this is good do not misunderstand me. There's a big difference between making 800K gross where 62.5% is illiquid and 800K net you can spend any time.

[edit] This isn't much different than getting a staff+ role at a MAANG, which includes liquidity, but of course upside depends in your view of the relative stock performance in the coming years. Also, this is well above median, we're probably talking top 5-10% of talent.

You don't think there's a market for this so called illiquid openai stock?
Potentially, of course, depending on a ton of factors. Stock agreements generally forbid transfers to third parties outside of like, death and divorce.

Also, they're probably RSUs at this point which these secondary market companies don't like because unlike options, they can't be exercised at the holder's discretion, or shares, which the employee owns outright, they don't vest until all conditions are met (which often means until liquidity event). If the conditions aren't met within 7 years the RSUs evaporate even if the company continues to do well.

So yeah you may be able to get a no-recourse loan or a forward sale, on bad terms, for the stock/cash to exercise the options. For RSUs you're either going to get even worse terms or politely told to come back once they vest.

Let me know if that's not the case though.

So. Couple of things. That is 800k tc and not base. Id say 250-300k based. But let us go with this being fully liquid and ignore vesting, exit event etc.

Half of that is tax - so 400k. A (crappy?) house in the bay area is 2-3M (sam I thought was pretty adamant about RTO). Property tax is 1.25% (so about 40k a year?). Currently at 5% we are looking at 150k a year on mortgage interest alone.

Private schooling is around 2-4k per kid (after tax). But let us say you are public schooling.

Assume you already have a car and no desire for fancy cars.

Assume you hardly eat out and don't go on vacations.

Assuming you have no other expenses id say yeah it is pretty easy to spend 800k!

I’ve lived the last year off of $120k post-tax dollars, and I’m only down to $65k. Saying that $800k is not hard to spend is unfathomable to me.

The exception is if you’re buying a house. That was a big kick in the teeth to the tune of $45k or so, after closing, due to house repairs and miscellaneous life stuff. But as far as starting a family goes, it’s worth it.

I don't get it. Did you buy a house for $45K, or was that the down payment and repairs?
The latter. The house was $260k.

I doubt we’ll pay it off any time soon, and it always felt silly saying you "own a house" when the bank owns it. I once asked what happens if you miss a payment after paying off 98%, and the closing lady said that the bank will repossess it.

Don't know how it works with local laws, but as far as I know, they could ,(doubt they would) repossess. Nevertheless, they would auction, take the 2% remaining, plus their admittedly ridiculous fees, and pay you back the rest. Not that it would end up a good deal, but it's not as disastrous as one would infer.
It would be a very bad deal because houses at auction don't fetch anywhere near their normal market value. They fetch their real value which is much lower than the value that the banks will happily allow you to borrow against (and the realtors charge their commission on).

Those very same banks who quite often happen to be close buddies with the parties that happen to have the winning bid at the auction. It's a very crooked game, make sure you know the rules and who the other players are.

> I once asked what happens if you miss a payment after paying off 98%, and the closing lady said that the bank will repossess it

In what country? In the US it's almost impossible for the bank to take your primary residence regardless of whether you stop paying when you owe 2% or 98%.

Anyway ownership is not the same as 100% equity. You still get market_value-loan after a shortsale.

I have chronic difficulty spending money, to the extent that I'm surprised if in any month my total outgoings including rent and utility bills exceeds €1k, but I also noticed a long time ago that my attitudes to money are unusual, and that most people want to spend slightly more than they earn regardless of income.

$800k/year, even if that were after tax, can be burned fast by sufficiently expensive hobbies like maintaining your own sufficiently large private plane or yacht. And if you want to go to orbit…

You’re lucky. My burn rate is $6k, and it makes me wince. Keep it up! (Or down, in this case.)

It’s not impossible to imagine burning $800k a year, but saying it’s "not hard" is so very Silicon Valley.

It's really not though. At least not the first couple of years.

Get any combination of a house, nice cars for you and spouse, a nice adventure vehicle, pay college debt and kids education and you're already in the red if you want top shelf stuff.

People who say "I couldn't even spend X amount of money" (where X is some large number) suffer from a lack of imagination. I could easily find ways to blow the world's entire GDP on bullshit if I somehow had the world's GDP to play with.
Most of these engineers are motivated by accomplishment. I very much doubt any of them would describe their work as soul draining.
I don't know, if I was financially independent, I'd do work that can be open sourced and free - get proper recognition - rather than giving it away to a corporation.
You're not exactly giving it away at 800K TC are you? :)
That work would be far removed from the industry forefront. With none of the resources and team with you. It takes millions to develop, train and run these AI models. Your plan sounds dreadfully lonely to me.
They could join Mistral AI, which has published weights for at least some of its models. Another option is Meta AI, which has published weights for Llama and Llama 2.
> dreadfully lonely

On your deathbed you are likely not going to look back favorably at all the years you put in for bigcorp advancing the industry. Even if you do have a few white papers published, so what? Effectively nobody will remember them in 10 years anyway. Trying to find meaning by working for others doesn't pan out for most people. Bank the cash, get out when you can.

Sounds nihilistic, no thanks
On my death bed I will only care about my kids and family. There is however hopefully a long time till I get there and during that time I would prefer to work on something awesome with good people. As we live in a capitalist system that is generally found in companies.
... because we have been working to this moment for often 10+ years.

People that were working on this stuff for a long time were not in it for money. It used to be a more artsy/philosophical field, and the money people moved on a long time ago for Enterprise edition Java or something.

On top of that, quitting now would feel like quitting a marathon during the final victory lap. If it turns out that there is still a significant gap between here and the end goal of AI research (which is different for different people), many people will retire. But lately, it has been feeling like we're on a final stretch.

> People that were working on this stuff for a long time were not in it for money.

It seems like 90% of OpenAI staff was going to join Sam Altman at Microsoft if the didn't return. Can you elaborate on that?

There are people out there who enjoy working on difficult problems and want to keep working on them.

I can't speak for 20 years from now, but if I had retire-now money,I would probably take a job with a small games company. Strangely enough, that's what I do now anyway.

> I would probably take a job with a small games company

This is the direction I am headed in. Once my boring SaaS/banking app money pile gets big enough, I'd like to get (back) into building new gaming experiences.

I've taken a slightly different approach, I'm just (planning on) working on the gaming experiences for my entire career, and not waiting for the "cash out to do what I love" moment.
I don't think I've ever heard ofthis being a thing lol

Why doesn't Jeff or Elon just stop working? People that make that much money don't work for the money.

Wait? Bezos works? I mean, apparently he works out a lot, but he's not CEO anymore. His main worries these days seem to revolve around his ridiculously large mega-yacht having to dock in container harbours due to its size and such...
> Why doesn't Jeff or Elon

They are extreme outliers, not really comparable to the general population.

lol. Do you know how much a "middle class" house costs in the bay area where these folks work? it is sort of crazy.
It's free market - if you think they're doing it wrong, you can win by doing what you think will work better.
Allegedly this is a problem Nvidia has encountered. Their share price increase means they have a lot of staff who are now rich through share ownership and who don't really need to put in the effort any more.

https://www.tomshardware.com/pc-components/gpus/wealthy-nvid...

It's not quite the same as OpenAI, but as soon as those engineers realise they're not having fun at work anymore, there's a risk they'll just leave or coast. For now they're probably riding high at the place to be, doing the "best" AI work there is. It could stay like that for a long time, but it probably won't last forever.

> why continue to do extremely difficult work?

For the same reason rich CEOs still go to work. They find pride in their work and it's a status symbol.

Imagine being part of the company that solved AGI.

Depending on the outcome that may cause you to walk proud or to go into hiding.
"If you've got enough money that you can't spend it in a lifetime"

Some people are modest, some people want space rockets.

Everyone here saying “smart people enjoy interesting work.” And that’s fine.

But let’s be honest, too. You dream about that first million. By the time you hit it, you want two…

By this logic we'd be running out of CEOs very fast.
This is making headlines recently but it's really not that outlandish of a pay level in tech in general (e.g. L7+ at Google make this kind of money, and it probably requires similar amounts of experience to reach this level), and seems to be working fine elsewhere.
It’s just the same as what happened with the banks after 2008. The majority of well earning people from the repackaged mortgage boom never got back to their previous earnings but a smaller number of people at the right place and the right time started making 5x.

This time it’s the low interest boom and software engineering.

Capitalism always finds a way to concentrate wealth.

Low interest? Interest rates are at 5.5%
Guess he was talking about last year when you had to pay the bank to have money on it
5.5% is still historically reasonable. Look at what rates were in the 70s and 80s. The issue is so many people got used to the ridiculously low rates over the last 10 or so years.
The issue is asset prices are higher now.
70s and 80s? You talking about the drastic fed action to break stagflation? How are you using that as a baseline?
> 5.5% is still historically reasonable

That doesn’t mean much on it’s, you can’t compare interest rates between different periods without taking into account inflation.

> were in the 70s and 80s

Economically that wasn’t a good period though.

This seems fairly unsurprising and entirely reasonable, given the compensations that have been paid to SWE in the area for fairly low impact and off-the-shelf work in recent years.